Guardrails
Guardrails are rules, filters, and policies that block unsafe inputs, limit risky outputs, and keep AI behavior aligned with product and brand standards.
They translate legal, security, and design requirements into what the product will and will not do, beyond a single disclaimer modal modal.
What it means
Guardrails include system prompts, classifiers, allow/deny lists, output validators, rate limits, and human review gates applied before or after generation.
Why designers should care
Users experience guardrails as refusals, redactions, tone shifts, and blocked actions. Copy and fallback UX must explain limits without feeling arbitrary.
Example
An image editor refuses NSFW prompts with specific guidance (“Try describing style without people”), and offers safe preset styles instead of a dead-end error.
Common mistakes
- • Generic “Something went wrong” when a policy block fires.
- • Inconsistent guardrails across channels (web vs API vs agent tools).
- • Guardrails that block legitimate power-user tasks with no override path for admins.